AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Complementary Domain Adaptation and Generalization for Unsupervised Continual Domain Shift Learning

Wonguk Cho, Jinha Park, Taesup Kim

ICCV 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 16 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
initialize the DA model from the previous DG model instead of the previous DA model (ablation)CoDAG · PACS (10 orders)TDA: 83.8 -> 87.6Improved
initialize the DA model from the previous DG model instead of the previous DA model (ablation)CoDAG · PACS (10 orders)All: 80.7 -> 82.9Improved
add SelNLPL noisy-label learning to the DG model (ablation)CoDAG · PACS pairs D0->D1 (no replay buffer)TDG avg: 73.0 -> 74.9Improved
add SelNLPL noisy-label learning to the DG model (ablation)CoDAG · PACS pairs D0->D1 (no replay buffer)FA avg: 89.9 -> 91.5Improved
add a replay buffer of size 200 for the DG model (ablation)CoDAG · PACS (10 orders)TDA, TDG, FAImproved
complementary two-model framework (DA and DG models exchanging initialization and pseudo-labels) versus best baseline RaTPCoDAG · PACS (10 orders)All: 79.7 -> 82.9Improved
CoDAG without replay buffer compared against baselines with replay buffer (ablation)CoDAG · PACS (10 orders)AllImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy